-
Story
-
Resolution: Won't Do
-
Normal
-
None
-
None
-
False
-
None
-
False
-
OCPSTRAT-751 - [Dev Preview] Extended Job Manager Operator
-
-
-
Workloads Sprint 244, Workloads Sprint 245, Workloads Sprint 246, Workloads Sprint 247, Workloads Sprint 248, Workloads Sprint 249, Workloads Sprint 250, Workloads Sprint 251, Workloads Sprint 252
Build a new operator that deploys the kueue solution from https://github.com/kubernetes-sigs/kueue.
Acceptance Criteria:
- the new operator is deploying the whole kueue stack
- the operator repo has a CSV definition (i.e. manifests directory)
- the operator has a CRD for the operator configuration
- the operator image can be built (through Dockerfile.rhel7)
The new operator is going to be a combination of https://github.com/openshift/cluster-resource-override-admission-operator and https://github.com/openshift/secondary-scheduler-operator.
TODO:
- when the operator is uninstalled, the underlying installed/owned resources are deleted (and all other secondary resources get deleted). One needs to properly set the owner reference to all secondary resources to point to the operator deployment.
- make the RBAC rules set as restricting as possible
- validate all dependent components are present (e.g. kube-scheduler)